Re: Re: skip entry with inherited tags

[Carsten Dominik <carsten.dominik@gmail.com>]

On Aug 4, 2010, at 11:38 PM, Martin Pohlack wrote:

> Hi Carsten,
>
> On 22.07.2010 09:38, Carsten Dominik wrote:
>> Hi Martin,
>>
>> On Jul 21, 2010, at 4:32 PM, Carsten Dominik wrote:
>>
>>> Hi Martin,
>>>
>>> I just looked at your patch.
>>>
>>> If I have a normal agenda (i.e. *not* a block agenda), then your
>>> patch will cause the preset filter *not* to be applied.
>>
>> OK, that was obviously incorrect, sorry about that.
>>
>> However, it still does not work correctly, here is the counter  
>> example:
>>
>> -----------------------------------------------------------------------------
>> * TODO test 1								  :a:
>> * TODO test 2								  :b:
>>
>> * at 1									  :a:
>>   SCHEDULED: <2010-07-22 Thu>
>> * at 1									  :b:
>>   SCHEDULED: <2010-07-22 Thu>
>> ------------------------------------------------------------------------------
>
> Here is what I found to be incorrect.
>
> * For unmodified org-mode, this ignores the org-agenda-filter-preset.
>  Is this what you mean or is anything else broken?
>
> * For my patched version, too much is hidden and only revealed after
>  clearing the global filter.  Did you mean anything else?
>
>> With this custom command:
>>
>> (setq org-agenda-custom-commands
>>       '(("x" "testmartin"
>> 	 ((agenda "" ((org-agenda-filter-preset '("+a"))))
>> 	  (alltodo "" ((org-agenda-filter-preset '("+b")))))
>> 	 nil nil)))
>>
>> the result is incorrect, both with and without your patch.
>
> But in different ways, as state above.

Yes, in different ways.

>
>> The internal logic of the filter and the preset filter is such that
>> it applies to the entire view, and you should not set in the local
>> options for a command that is part of a block agenda view.
>
> Well, it is already partly there in that local filters are stored in
> text properties for each line.  Maybe we can extend this a bit to  
> remove
> this limitation?

I think this will be tidious.  For block agendas, I guess a skip
condition will be better than filtering.

Best wishes

- Carsten

>
> I will look into it if I have more time.
>
> Cheers,
> Martin

- Carsten